Developed and delivered real-time STTv2 streaming integration for Scribe v2 within the livekit/agents repository, enabling streaming speech-to-text with voice activity detection and improved session management. Leveraged Python and async programming to reduce perceived latency and provide smoother user interactions during real-time transcription. Focused on API integration and real-time processing, the work introduced streaming capabilities that enhanced the responsiveness and reliability of Scribe v2’s speech recognition features. Collaborated cross-functionally to align technical implementation with product requirements, ensuring the integration addressed both performance and usability needs. The project demonstrated depth in asynchronous systems and real-time audio processing using modern Python techniques.